Table 1a.

Summary of model metrics.

	Metrics	Minimum	5th Percentile	25th Percentile	Median	75th Percentile	95th Percentile	Maximum	Mean	SD	Range
XGBoost	Accuracy	0.684	0.741	0.766	0.790	0.806	0.836	0.898	0.789	0.026	0.210
	F1	0.686	0.750	0.774	0.784	0.810	0.835	0.897	0.787	0.031	0.204
	Sensitivity	0.680	0.757	0.790	0.806	0.821	0.853	0.901	0.802	0.026	0.224
	Specificity	0.592	0.708	0.749	0.786	0.815	0.852	0.947	0.789	0.037	0.348
	PPV	0.680	0.761	0.787	0.818	0.847	0.884	0.958	0.818	0.035	0.273
	NPV	0.567	0.676	0.722	0.753	0.785	0.829	0.930	0.761	0.046	0.354
	AUROC	0.772	0.831	0.856	0.867	0.884	0.903	0.948	0.868	0.025	0.171
Random Forest	Accuracy	0.675	0.729	0.771	0.778	0.801	0.812	0.892	0.784	0.027	0.224
	F1	0.687	0.740	0.771	0.776	0.809	0.816	0.884	0.785	0.030	0.201
	Sensitivity	0.665	0.745	0.782	0.799	0.804	0.847	0.895	0.793	0.024	0.229
	Specificity	0.584	0.709	0.748	0.784	0.803	0.845	0.927	0.771	0.041	0.340
	PPV	0.676	0.740	0.779	0.813	0.846	0.857	0.948	0.810	0.045	0.270
	NPV	0.555	0.661	0.720	0.736	0.772	0.826	0.908	0.750	0.045	0.359
	AUROC	0.757	0.824	0.842	0.860	0.887	0.900	0.928	0.857	0.022	0.175
Artificial Neural Network	Accuracy	0.689	0.736	0.761	0.786	0.805	0.830	0.877	0.781	0.021	0.194
	F1	0.677	0.732	0.750	0.783	0.790	0.818	0.888	0.774	0.027	0.211
	Sensitivity	0.672	0.749	0.779	0.794	0.802	0.834	0.884	0.796	0.021	0.216
	Specificity	0.591	0.707	0.749	0.768	0.799	0.835	0.928	0.768	0.035	0.327
	PPV	0.659	0.748	0.780	0.809	0.835	0.859	0.940	0.808	0.029	0.274
	NPV	0.550	0.665	0.718	0.752	0.772	0.817	0.912	0.749	0.047	0.361
	AUROC	0.751	0.821	0.839	0.866	0.882	0.891	0.949	0.847	0.027	0.192
Adaptive Boosting	Accuracy	0.683	0.731	0.761	0.790	0.793	0.821	0.885	0.775	0.023	0.199
	F1	0.674	0.739	0.760	0.774	0.801	0.828	0.890	0.775	0.029	0.224
	Sensitivity	0.671	0.753	0.783	0.811	0.809	0.839	0.889	0.797	0.019	0.216
	Specificity	0.585	0.694	0.746	0.777	0.803	0.853	0.941	0.772	0.045	0.354
	PPV	0.676	0.742	0.772	0.805	0.843	0.861	0.951	0.817	0.045	0.277
	NPV	0.567	0.664	0.717	0.751	0.784	0.825	0.929	0.750	0.047	0.358
	AUROC	0.755	0.815	0.840	0.860	0.865	0.894	0.929	0.862	0.025	0.175

Note: Summary of model metrics within the test set for each of the four machine learning techniques (XGBoost, Random Forest, Artificial Neural Network, and Adaptive Boosting) based upon bootstrap simulation.

Abbreviations: AUROC, area under the receiver operator characteristic curve; NPV, negative predictive value; PPV, positive predictive value.